User and Password for RShiny

Can we maybe have a user and password for a shiny application running openEO processes?

  • Short description: I need to have a general access to the rshiny so it can be publishable
  • Specific request/idea/comment: I’m developing a RShiny program where a user can interact with S5P data. For now, it’s necessary for the user to either have the code (through git clone or something like that), but ideally it would be nice to publish the UI as a real rshiny program. For that, I need a general user and password so people can access without requiring to git clone or pass their own credentials (and have them).

Do you believe this to be feasible somehow?

Thank you in advance :slight_smile:

I don’t fully understand: What service do you need a username/password for? For openEO Platform? For GitHub? For something R-related?

For the openEO Platform. Something like the Early Approval one for example.

It’s an issue which has been already reported, like here Non-interactive authentication

anyway, @stefaan.lippens wrote this nice explanation in the Python Client docs about this scenario:
https://open-eo.github.io/openeo-python-client/auth.html#authentication-for-long-running-applications-and-non-interactive-contexts

Still, there are two issues:

  1. There’s still an interactive step at the beginning, which makes this mostly unsuitable for auto-deployments
  2. It is only (easily) available in Python as of now

So a better solution is still to be found.

The refresh tokens can be shipped anywhere, keeping them secure is the hard part imo, but there might be other solutions for that (maybe https://www.vaultproject.io/).
To obtain and mange tokens from the command line, there’s: Introduction - oidc-agent
That might help users of other programming languages where the client doesn’t have full-featured oidc management built in.

We’ll need to discuss on what type of account is appropriate for this, but perhaps it’s best to figure out technical details first? (I’m guessing it’s low usage anyway?)

Yes, the usage would be low as even shiny limits the number of users anyway